Visual Hull Refinement with Hierarchical Clustering Deformation for 3-D Reconstruction
Yung-Yang Chiang, Huei‐Yung Lin, Min-Liang Wang · 2021
This paper presents a hierarchical clustering deformation approach to improve the visual hull based 3-D model re-construction technique. The proposed technique takes a reference 3-D computer model and multi-view image captures of the target as inputs. A model deformation process is then carried out under the constraints imposed by the silhouettes of acquired images and the camera geometry. Therefore, it is able to approximate the true 3-D models with limited images. To facilitate this constrained 3-D modeling, a clustering method is proposed to divide the 3-D model into several sub-groups for individual matching and deformation. In the experiments, the real images of several test objects are used for our performance evaluation. The results have demonstrated the feasibility of the proposed hierarchical deformation technique for 3-D model reconstruction.